Llamacpp imatrix Quantizations of xLAM-8x7b-r

Using llama.cpp release b3634 for quantization.

Original model: https://huggingface.co/Salesforce/xLAM-8x7b-r

All quants made using imatrix option with dataset from here

Run them in LM Studio

Prompt format

<s> [INST] {system_prompt}

{prompt} [/INST] 

Download a file (not the whole branch) from below:

Filename Quant type File Size Split Description
xLAM-8x7b-r-f16.gguf f16 93.41GB true Full F16 weights.
xLAM-8x7b-r-Q8_0.gguf Q8_0 49.63GB false Extremely high quality, generally unneeded but max available quant.
xLAM-8x7b-r-Q6_K_L.gguf Q6_K_L 38.44GB false Uses Q8_0 for embed and output weights. Very high quality, near perfect, recommended.
xLAM-8x7b-r-Q6_K.gguf Q6_K 38.38GB false Very high quality, near perfect, recommended.
xLAM-8x7b-r-Q5_K_L.gguf Q5_K_L 33.31GB false Uses Q8_0 for embed and output weights. High quality, recommended.
xLAM-8x7b-r-Q5_K_M.gguf Q5_K_M 33.23GB false High quality, recommended.
xLAM-8x7b-r-Q5_K_S.gguf Q5_K_S 32.23GB false High quality, recommended.
xLAM-8x7b-r-Q4_K_L.gguf Q4_K_L 28.55GB false Uses Q8_0 for embed and output weights. Good quality, recommended.
xLAM-8x7b-r-Q4_K_M.gguf Q4_K_M 28.45GB false Good quality, default size for must use cases, recommended.
xLAM-8x7b-r-Q4_K_S.gguf Q4_K_S 26.75GB false Slightly lower quality with more space savings, recommended.
xLAM-8x7b-r-Q4_0.gguf Q4_0 26.56GB false Legacy format, generally not worth using over similarly sized formats
xLAM-8x7b-r-Q4_0_8_8.gguf Q4_0_8_8 26.44GB false Optimized for ARM and CPU inference, much faster than Q4_0 at similar quality.
xLAM-8x7b-r-Q4_0_4_8.gguf Q4_0_4_8 26.44GB false Optimized for ARM and CPU inference, much faster than Q4_0 at similar quality.
xLAM-8x7b-r-Q4_0_4_4.gguf Q4_0_4_4 26.44GB false Optimized for ARM and CPU inference, much faster than Q4_0 at similar quality.
xLAM-8x7b-r-IQ4_XS.gguf IQ4_XS 25.08GB false Decent quality, smaller than Q4_K_S with similar performance, recommended.
xLAM-8x7b-r-Q3_K_XL.gguf Q3_K_XL 24.28GB false Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability.
xLAM-8x7b-r-Q3_K_L.gguf Q3_K_L 24.17GB false Lower quality but usable, good for low RAM availability.
xLAM-8x7b-r-Q3_K_M.gguf Q3_K_M 22.55GB false Low quality.
xLAM-8x7b-r-IQ3_M.gguf IQ3_M 21.43GB false Medium-low quality, new method with decent performance comparable to Q3_K_M.
xLAM-8x7b-r-Q3_K_S.gguf Q3_K_S 20.43GB false Low quality, not recommended.
xLAM-8x7b-r-IQ3_XS.gguf IQ3_XS 19.35GB false Lower quality, new method with decent performance, slightly better than Q3_K_S.
xLAM-8x7b-r-Q2_K_L.gguf Q2_K_L 17.44GB false Uses Q8_0 for embed and output weights. Very low quality but surprisingly usable.
xLAM-8x7b-r-Q2_K.gguf Q2_K 17.31GB false Very low quality but surprisingly usable.
xLAM-8x7b-r-IQ2_M.gguf IQ2_M 15.50GB false Relatively low quality, uses SOTA techniques to be surprisingly usable.
xLAM-8x7b-r-IQ2_S.gguf IQ2_S 14.13GB false Low quality, uses SOTA techniques to be usable.

Embed/output weights

Some of these quants (Q3_K_XL, Q4_K_L etc) are the standard quantization method with the embeddings and output weights quantized to Q8_0 instead of what they would normally default to.

Some say that this improves the quality, others don't notice any difference. If you use these models PLEASE COMMENT with your findings. I would like feedback that these are actually used and useful so I don't keep uploading quants no one is using.

Thanks!

Credits

Thank you kalomaze and Dampf for assistance in creating the imatrix calibration dataset

Thank you ZeroWw for the inspiration to experiment with embed/output

Downloading using huggingface-cli

First, make sure you have hugginface-cli installed:

pip install -U "huggingface_hub[cli]"

Then, you can target the specific file you want:

huggingface-cli download bartowski/xLAM-8x7b-r-GGUF --include "xLAM-8x7b-r-Q4_K_M.gguf" --local-dir ./

If the model is bigger than 50GB, it will have been split into multiple files. In order to download them all to a local folder, run:

huggingface-cli download bartowski/xLAM-8x7b-r-GGUF --include "xLAM-8x7b-r-Q8_0/*" --local-dir ./

You can either specify a new local-dir (xLAM-8x7b-r-Q8_0) or download them all in place (./)

Which file should I choose?

A great write up with charts showing various performances is provided by Artefact2 here

The first thing to figure out is how big a model you can run. To do this, you'll need to figure out how much RAM and/or VRAM you have.

If you want your model running as FAST as possible, you'll want to fit the whole thing on your GPU's VRAM. Aim for a quant with a file size 1-2GB smaller than your GPU's total VRAM.

If you want the absolute maximum quality, add both your system RAM and your GPU's VRAM together, then similarly grab a quant with a file size 1-2GB Smaller than that total.

Next, you'll need to decide if you want to use an 'I-quant' or a 'K-quant'.

If you don't want to think too much, grab one of the K-quants. These are in format 'QX_K_X', like Q5_K_M.

If you want to get more into the weeds, you can check out this extremely useful feature chart:

llama.cpp feature matrix

But basically, if you're aiming for below Q4, and you're running cuBLAS (Nvidia) or rocBLAS (AMD), you should look towards the I-quants. These are in format IQX_X, like IQ3_M. These are newer and offer better performance for their size.

These I-quants can also be used on CPU and Apple Metal, but will be slower than their K-quant equivalent, so speed vs performance is a tradeoff you'll have to decide.

The I-quants are not compatible with Vulcan, which is also AMD, so if you have an AMD card double check if you're using the rocBLAS build or the Vulcan build. At the time of writing this, LM Studio has a preview with ROCm support, and other inference engines have specific builds for ROCm.

Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski

Downloads last month
327
GGUF
Model size
46.7B params
Architecture
llama

2-bit

3-bit

4-bit

5-bit

6-bit

8-bit

16-bit

Inference Examples
Unable to determine this model's library. Check the docs .

Model tree for bartowski/xLAM-8x7b-r-GGUF

Quantized
(8)
this model

Dataset used to train bartowski/xLAM-8x7b-r-GGUF